Latest posts
-
How I Built This Site: Dark WordPress from Scratch — No Page Builder
A full behind-the-scenes breakdown of how backbazaar.com was built — the CSS system, the WooCommerce dark skin, the block editor workarounds, and everything that broke along the way.
-
How I Built This Site: A Dark WordPress from Scratch
A full behind-the-scenes breakdown of how backbazaar.com was built — the CSS design system, WooCommerce dark skin, block editor workarounds, and everything that broke.
-
AI Agents in Production: What Breaks and What Doesn’t
Agentic AI systems are impressive in demos and brittle in production. Here’s a practical breakdown of the failure modes, and the architectural patterns that actually hold up.
-
NRI Financial Planning in the Gulf: What the Generic Guides Miss
Most NRI financial planning guides are written for Indian-Americans. If you’re in Saudi Arabia, UAE, or Qatar, the rules are different — and the mistakes are expensive.
-
The Honest Guide to LLM Evaluation
Evaluating LLMs is harder than evaluating any other ML system. Here’s a practical framework — what to measure, how to measure it, and which shortcuts to avoid.
-
Building Anomaly Detection on Streaming Telecom Data
How we built a real-time anomaly detection system on Kafka + PySpark Structured Streaming to catch network and billing issues within minutes instead of the next morning.
-
Why Most AI Projects Fail Before They Ship
The technical part is rarely what kills an AI project. Here’s the honest list of what actually goes wrong — and how to fix each one before it’s too late.
-
5 PySpark Performance Patterns That Actually Matter at Scale
Most PySpark optimisation guides are theoretical. These five patterns came from real 100M+ record pipelines running under production pressure in telecom.
-
RAG Systems in Production: What the Tutorials Don’t Tell You
Retrieval-Augmented Generation looks elegant in demos. In production, it breaks in ways that are subtle and expensive. Here’s what I’ve learned building RAG over real enterprise documents.
-
Building a Churn Model That Actually Gets Used
Most ML projects die in the gap between ‘the model works’ and ‘the business uses it’. Here’s how we crossed that gap.